Disaster Test

Shut down all hosts on the source side and shut down the source array. If shutting down the primary array is not feasible, disconnect the replication link between the two arrays while simultaneously shutting down the hosts; this action mimics a disaster scenario to the secondary side.

In a replicated data cluster, the service group fails over to hostc or hostd if all devices were originally in the PAIR state, that is, no synchronization was in progress at the time of disaster.

In a global cluster, the administrator is notified of the failure. The administrator can then initiate the failover by declaring an outage.